Estimation of neutral sugars and sugar alcohols in biological fluids by gas-liquid chromatography

Jo.urz& of Chromztogr~&, 184 (1980) 457-470 Chroanuto&Ta&lhk Revikvs EIsevier Scientic Prrb!ishingCompany, Amsterdam-

printed in The NetExrhnds

140

(=HREv.

ESTIMATION BIOLOGICAL

OF NEUTRAL SUGARS AND SUGAR ALCOHOLS FLUIDS BY GAS-LIQUID CHROMATOGRAPHY

M

M. F. LAKER’ Depprrmentof Lm&n

Chemica! Patkohgy SE1 7EH (Great Britain)

md Metakolic

Disorders, St. TkonmA

Hospit& Medica

Sckoo&

- _ .._.._

_ 457

(Received May Zznd, 1980)

coNTJ3Ts 1. Introduction

. . . _ _ . . _ . . . ..-

-.....-

-..-

formation- _ _ _ _ _ _ _ - - __-.____________..___ 2.1. Methyl derivatives. . _ . . _ _ 2.2. Acyl derivatives_ . . . . _ . . 22.1. Acetyl derivatives _ . . . 22.2. Trifluoroacetyl derivatives. 2.3. TrimethyIsilylderivatives_ _ _ _ 2.4. n-Butylboronate derivatives_ _ . 2.5. Comparison of derivatives . . . Columns and stationary phases _ . _ Qualitative and quantitativeanalysis _ 4.1. Qualitative analysis _ . _ . . _ 4.2. Quantitation . . . . . . . . . 4.3. Detailed analytical procedure _ . Ferfomlanceofassays _ _ _ _ _ . _ Conclusions. . _ . . _ _ . _ _ . _ Acknowledgements _ . _ . _ . _ .

2. Derivative

3.

4.

5. 6.

7.

s.summuy.

_ - _ _ -_

458 45s 45s 45s 460 460 461 462 462 462 462 463 466 467 468 468

_ - _ _ - _

References _ _ . _ _ . . . . _ . . _

1. XNTRODUCX’ION

Many sugars and sugar alcohols occur natuially in biological fluids, often in low concentrations. Since more than 40 such compounds have been demonstrated in human urine1 and 21 in human seminal plasma l, it is necessary to use specific and sensitive techniques for their detection and measurement. Gas-liquid chromatography (GLC) is particularly suitable for such studies since it combines great sensitivity with high resolution. The purpose of the present review is to survey the use of GLC for the measurement of free sugars and sugar akohols in biological fluids. Its use in the structural analysis of the carbohydrate components of glycolipids and glycoproteins has been considered in previous review$-5.

* Present address: Depastment of Clinical Biochemistry, Royal Victoria Infirmary, Newcastle upon Tyue NE1 4LP, Great Biitain. 037k4435slgo/aoclcMooo JS02.80 0 1980 Ekevier ScientificPublishing Company

M.F.LAJU5.R

458 2. DERIVATIV!E

FORMATiON

Sugars and sugar alcohols are non-volatile compounds which cannot be analysed by GLC unless volatile derivatives are first formed. Many derivatisation procedures have been described which involve methylation, acylation, trimethylsiiylation or butylboronate formation. In all these procedures several chromatographic peaks may result from a sir&e sugar due to CL-and /?-anomers and ring isomers occurring, each of which may form a separate derivatives. Thus analysis of samples containing seve& sugars may produce a very complex chromatogram. Patterns characteristic of an individual sugar may be an advantage when investigaing the composition of complex saccharides such as glycoproteins, since their recognition wilI increase the degree of confidence with which a single sugar may be identiEecF. However, the formation of multiple derivatives has several disadvantages in quantitative analysis of free sugars. Sensitivity will be reduced, since the amount needed to give a minimum detectable response will be greater if several rather than single peaks are formed. Chromatograms may be crowded, with ovedapping and coincidence of peaks. In addition, inaccuracies would resuli if variations occurred in the proportions of each isomer formed within an analytical series. To overcome these problems, additional derivatisations have been described which are used in conjunction with acetylation and trimethyIsiIyM.ion in order to reduce the incidence of muhiple products. These additional procedures are alditol, oxime, methyloxime and aldononitri!e formation. 2.1. Methy

derivarives

Before the introduction of GLC sugars were separated as methylglycosides. It is therefore not surprising that in the first report of GLC of sugars’ anomers of D-xylose, L-arabinose, ~-glucose, D-maznose and r+galactose were separated as methyl derivatives. The anomeric forms of a single monosaccharide have also been resolved using this technique*_ Kirche? also showed that derivatised sucrose was sticiectly volatile for chromatography and in 1962 Bishop9 demonstrated that mixtures of’fully methylated disaccharides could be resolved. ,Methylation has also been used to separate isomeric sugar aIcohoW”. Early methods of methylation were cumbersome, but simplified procedures ‘have now been described which are more suitable for the analysis of large numbers of small specimenP_ This type of derivatisation, in which sugars are methylated by methylsulphinylmethide carbion generated from sodium hydride and dimethyl sulpho_xide has been used in the separation and quantitation of disaccharides occurring in f-1 homogenates=_ 2.2. AcyZ derivatives 2.2-Z. AcetyZ derivatives. The separation of sugar alcohols and monosaccharides as acetyl derivatives was first described by Gunner et al.= who, in addition to separating sugar alcohols of diKering moiecular weight, found that isomeric polyols could be resolved. Anomeric glycose acetates gave separate peaks, and derivatives of epimers had different retention times. The technique was extended to the separation of disaccharides by using more thermostable stationary phases which were coated in thinner f& on the inert supporP~r5.

GLC OF SUGARS AND SUGAR ALCOEZOLS

459

To eliminate the formationof multiple derivativesfrom a single sugar the carbonyl group of monosaccharidescan he reduced with sodium borohydride to form the correspondingsugar alcoholsIs. These are then acetylated before being injectedonto the chromatograph.This techniquehas been extendedto the analysis of complex carbohydratesfrom soil samplesby Oades” who describedderivatisation proceduressuitable for the routine handling of specimens.Samples were dried and acetylatedby adding acetic anhydrideand pyridine(50:50, v/v), then &nding overnight at room temperature.This procedure was adapted to the analysis of sugar alcohols in the serum and urine of uraemic and diabetic patients by PitkBnerY8, sodium borohydridereductionbeing omitted. Since glucose was not separatedfrom galactitol or sorbitol it was removed by initialincubationof the sampiewith glucose oxidase, the resultiaggluconatebeing adsorbedwith an ion-exchangeresin.This type of procedurehas been used to study sugar alcohols in cerebrospinalfluidlg*‘Oand adaptedto quantifypolyols in human lens tissuefrom a patientwith galactosaemiazl. Reduction followed by acetylation has been used to quantify disaccharidesin biologicalfluidsz2.There are howeverdisadvantagesto this type of method being used to estimate monosaccharidelevels: naturally occurring polyols would cause errors in quantitation,and two monosaccharidesmay give rise to the same sugar alcohol (e.g., arabinoseand lyxose) An alternative approach is the preliminary conversion of aldoses to the correspondingmethyloximeor aldononitrilederivative,followedby acetylationof free hydroxylgroups. The incidenceof multiplederivativesfrom a sin&e sugar is reduced since a//Lanomerisationdoes not occur. This is becausethese additionalprocedures involve reactionswith the C-l position of aldoses. Sincethey are fUrlyreduced,sugar alcohols do not form methyloximeor aldononitrilederivatives.Derivatisationwith methoxylaminehydrochloride,followed by acetylation,has been used by Murphy and Pennock” to measuremonosaccharidesin blood and urine.Singlepeakswereobtained for ghrcose,fructose,mannoseand xylose, but o’&ermonosaccharidesgave two peaks, presumablydue to formation of syn- and anti-isomers.One galactosepeak coincided with that of glucose. A simi!arprocedurehas been used for disaccharideestimationby Schwindet aLz’, one peak being obtained for maltose, but two resultingfrom lactose. Allen et QZ.~*,measuringgalactitol levels in liver and amniotic fluid, used the derivatisationproceduredescribedby Murphy and Pennockz to modify the retentiontime of glucose and thus preventit interfering. Polyacetyl aldononitrile derivatives may be synthesised from aldoses by dissolving sugars in pyridine, adding hydroxylamine hydrochloride, and heating. Acetic anhydrideis added, convertingsugar oximes which have been formed to the correspondingaldononitriles,and acetylatingfree hydroxyl group~~*‘~. Polyols do not react with hydroxylaminehydrochloride,and thus form fully acetylatedderivatives. This type of procedure has been used by Szafranek et ~1.~’ to separbtc 32 aldoses and polyols on glass capillarycolumns. The same group later separatedand quantified17 sugars and polyols occurringin human urine2s,and adapted the technique to the analysis of lens tissue from cataractous subjectszg.A method for measuringn-mannose in serum as its aldononitrileacetatehas been described=, and this type of derivatisation’hasalso been used to estimate sugar and sugar alcohol concentrationsin bacteriologicalculture media31. Samplesderivatiscdby the variousacetylationproceduresdescribedhave been

460

M.F.JLAXER

prepared for inje&on in severa! ways. Seine authors have injected tie reaction products directly onto the chromatography coIum~P*~~~‘, while others have first evaporated the reaction mixture and then dissolved the residue in a small volume of solvent, e.g., chloroform 17-, trich!oromethanP or acetone3’, prior to injection. 2.2.2. Trz~zzuruacety~derkzztives. Derivatisation of sugars with Mluoroacetic anhydride to form the corresponding esters was original!y described by Bourne et ~i?~_ Vi!kas et ai,u resolved monosaccharides, disaccharides and trisaccharides as their trit!uoroacetates, and showed that for ghcose this derivative was more volatile than the corresponding trimethylsilyl ether. Trifluoroacetyl derivatives have been used for separating cyclical sugar alcohok?, combined with alditol formation to measure sugars in- plasma35and combined with methyloxime synthesis for disaccharide and trisaccharide analysis”. Severa! different trifhzoroacetylation procedures have been described. Vi!kas

et al.= prepared derivatives by using trifluoroacetic anhydride and sodium trifluoroacetate in acetonitrile. Trifluoroacetic anhydride has also been used with pyridinP, tetrahydrofurarP and dichloromethantii as solvents. More reproducible derivatisations have been claimed if N-methyl-bis(trifluoroacetamide) is used, with an equal volume of pyridine as solver@. 2.3. T’inzethyIsiiyZ derivatives Trimethylsilyl (TIMS) derivatives have been widely used for the analysis of sugar mixtures since Sweeley et ai. 3g described a simple method of preparation. A mixture of hexamethyldisila (HMDS), trime’&ylchlorosilane (TMCS) and anhydrous pyridine was added to dried su_m, derivatisation occurring rapidly at room temperature with virtually quantitative yields. Good separations of various

pentoses and hexoses we& obtained although some isomeric polyols (particularly galactitol, mannitol and sorbitol) were not resolved. A fine precipitate of ammonium chloride was formed in the reaction mixture, and since this was injected directly, some contamination of the chromatographic column occurred. This procedure has been adapted to the measurement of monosaccharides and sugar alcohols in blood, urine and tissues by Wells and co-workersm2, p&s being quantified by direct comparison with the appropriate standard. A similar derivatisation procedure has been used by Bhatti and clamp43 who modified the technique to include internal standardisation and temperature programming, mea-

suring simultaneously urinary mono- and disaccharides. Ammonium chloride conramination of the column was avoided by De Neef& who extracted the products of derivatisation with hexane before injection. Sweeley-type derivatisations have been applied, with some modification, to the measurement of sugar alcohols in tiss~_&~, blood, cerebrospinal fluid and seminal fluid;” and urine50,and to the estimation of sugars in cerebrospinal ffuid5’. The T!W3 ethers of mannitol, sorbitol and gala&to!, and those of isomeric pentitols are not as well resokd as the corresponding acetyl derivatives. Some authors report results as total hexitols and pentitolssr, while others include additional methods to differentiate individual sugar alcohols. These additional techniques include enzymatic and microbiological procedures~, and GLC methods which include

GLC OF SUCMRS

AND

SUGAR

ALCOHOLS

46L

acetylation in addition to trimetbylsilylation 54. Partial separation of mannitol from sorbitol and galactitol has heen reported by using 2.7-m columns with a high percentage, moderately selective stationary phase (10% EW-27; ref. 55). TMShexitols and also TMS-pentitols have been resolved with capillary columns2, TMS ethers are susceptible to hydrolysis if traces of water OF acid are present. This possibility may be reduced if more powerful silyl donors than HMDS are used, such as N,O-bis(trimethyisilyl)acetamide (BSA), N,O-bis(trimethyisilyl)triffuoroacetamide (BSTFA), N-trirnethylsilyiimidazo~e (TM%), OF N-methyl-N-trimethylsilyitrifluoroacetamide (MSTFA) %e5’. TMSI has been used to derivatise sugars in urine, mW9, erythrocyte~~~ and cerebrospinal fluid61. Procedures using BSA, with TMCS as catalyst, have been described in which sugar alcohols, monosaccharides and disaccharides in plasma, urine and bile have heen determined55B62-63_ Although in theory BSTFA should be a self-catalyzing silyl donor, it has been found necessary to add TMCS as an additional catalyst for rapid quantitative derivatisation56*62. This combination of reagents has been used in methods to measure monosaccharides and poiyois in blood and urinea*65 and in the study of oligosaccharides in milkG. Pyridine is the solvent used for the majority of studies involving silyiation although other solvents, such as dimethylformamide and dimethyisuIphoxide, have also been ~se@‘-~~. Among the advantages cited for dime*Jylformamide compared to pyridine are : (I) when used with 2-hydroxypyridine as a catalyst, a more reproducible molar response factor is obtained, particularly for N-ace@ sugars; (2) it is more pleasant to work with; (3) it is more volatile, and causes less tailing of the solvent front. Additional derivatisation procedures have been combined with trimethylsiiyiation to decrease the incidence of multiple derivatives. Combined alditol-TMS derivatives have not been used, since these compounds are less satisfactorily resolved than the corresponding acetyl esters. Sweeiey et ~1.“~ described methods for the preparation of oxime-TMS sugars, and this procedure has been adapted to measure glucose in urine52. Since the retention times of monosaccharides are modified, additional separations may be achieved by this method, e.g. @-galactose and galactitol could not be resolved as TMS ethers, but were separated when the oxime-TMS derivative of galactose was formed 46. The synthesis of methyloximc-TMS sugars has been applied to the quantitation of monosaccharides in plasma and urineag6’ and seminal fluid?. The formation of oxime and methyloxime derivatives prior to si!yiation reduces the number of multiple derivatives. However, two peaks may result from a La&one single suga$6-71 presumably due to the formation of syn- and a&isomers. and methylglycoside formation have heen combined with trimethyisiiylation but result in little reduction in the number of derivatives4*5*39. 2.4. n-Butylboromzte derivatives

Butyiboronate derivatives have been shown to give useful separations of mannitol, sorbitol and galactitol ‘4 the derivatives being stable in the presence of water, As with other derivatives, multiple peaks were obtained from aldoses. One of those from glucose had the same retention time as mannitoi. This derivatisation has been used as the basis for the quantitation of sugar alcohols in pharmaceutical

462

M. F. LAJKER

preparation~~and has recentlybeen adapted to the analysis of mannitol in urine74. In this method interferencedue to glucose was overcome by first reducing it to sorbitof. 25.

of &J-z-twz-~-es

coln~on

Of the derivativesdescribed, acetyl esters and TMS ethers have been most widely used for measuring sugar and sugar alcohol concentrations. The major adiran~ for a&y1 esters are greater stability to water, and better resolution of isomeric polyols. TMS ethers are more volatile, a property of great advantage in the estimation of di- and trisaaharides. Methyl&ion techniqueshave not been as widely used for measuringfree sugar and polyol concentrations,probably because early methods of preparationwere cumbersome. Methyl derivativesare more stable to water than TMS ethersand more volatile than acetyl esters.TriEuoroacet.ates are detectabIeat lower concentrationsnsing an electron-capturedetectorthan derivatives which are quantified with a flame-ionisation detector. Using trifluoroacetylation. Pritchardand Niedermeier75 wereable to quantitatethe carbohydratecompositionof 0.1 pg glycoprotein. Disadvantages of the technique are that the derivativesare extreme!y sussptible to hydrolysis, and some authors have reported unreliable quantitativeresulting.The principleadvantagefor n-butylboronatederivativesis that they resisthydrolysisby water. 3. COLUMNS

AND

STATIONARY

PHASES

Stationary phases tbat have been used are presentedin Table 1. Non-polar liquid phases have generallyken used for separatingTMS derivativeswhile acetyl esters have been resolved using more polar phases. The ratios of liquid to inert support have varied from 1 to 25 % (w/w) although in generallow percentageshave been used. Gas-Chrom P, Gas-Chrom Q and Chromosorb W have alI proved satisfactory as inert supports. 4. QUAJXI-ATIVE

AND

QUANTITAlTVE

ANALYSIS

4. I. Quditativc analysti Identificationand characterisationof sugars and sugar alcohols by GLC is dependenton several factors, e.g. the ability to separate differentderivatives,the stationary phase used, the carriergas flow-rate, and the column temperature.The latter two are difficultto reproduceprecisely, but the effect of variations may be minimisedby definingthe position of peaks relative to marker substances_This is done by determiningreIativeretentiontimes or methyleneunit values. The relative retentiontime is the ratio of the time for the test sugar to appearrelativeto that for a reference internal standard compound. Fluctuations in column temperatureand carriergas flow-rate would afEct the test and internal standard retentiontimes in proportion, and the relative retentiontime would not alter. Thus i&ntification of sugars relativeto the position of a marher peak is subject to less error than being dependenton an absolute retentiontime. Extensivelists of relativeretentiontimes

GLC

OF SUGARS

AND

SUGAR

ALCOHOLS

463

using several stationary phases have been published for different deivatives of monosaccharides, disaccharides and sugar alcohols80~81, for TMS-carbohydrates39, for TMSdisaa3&des62, for alditol acetates16-*7and for oxime- and methyioxime-TMSmonosaccharideP_ Characterisation by methylene unit values involves determining the position of each peak relative to two straight chain hydrocarbon standards, one with a shorter and the other with a longer retention time than the test peak, e_g_ a sugar with a methylene unit value of 18.5 would appear half way between octadecane and nonadecane. This method has been used to characterise separations of methyl ethers of disaccharides*2, aldononitrile acetates of monosaccharides and polyols27-29, and TMS derivatives of mono-, di- and trisaccharideP. Both methods use the time of the maximum peak height. Whether useful separations of two sugars with similar retention times will be achieved depends not only on the retention time but also on the width of the peaks (Le. on the efficiency of the column). The column efficiency may vary considerably with different batches of stationary phase. 4.2. Qumtitation Quantitative results are usually obtained by using internal standardisation. A known amount of internal standard is added to the test samples and standard solutions at the start of the analytical procedure. The detector response ratio or molar response factor for each sugar of interest is determined relative to that of the internal standard. This must be calculated for each sugar since response factors cannot be predicted, because although the response shou!d be proportional to the mass of the substance, detectors do not “see” all chenncal groups equally well. This applies particularly to TMS derivatives8’. Using the response factors obtained from standard solutions, the concentrations of sugars in test samples may be determined by measuring the ratio of test to internal standard_ Using an internal standard has three main advantages : (1) identification of test peaks is more certain; (2) samples are injected with a microlitre syringe with which it is difficult to deliver a precise volume. With internal standardisation a ratio is measured and thus fluctuations in volume may be compensated for, and (3) volatile solvents may be used and some evaporation of these may occur. With internal standardisation these losses are corrected. An “ideal” internal standard should have certain properties: (a) it should be chemically similar to the substances being estimated; (b) it should give a single chromatographic peak which is well separated from other components; (c) it should be stable; (d) it should be available in a highly purified form and (e) it should not occur naturally in the fluid being analyzedTwo groups of substances have been used as internal standards for sugar analyses, other sugars and hydrocarbons. Both give reference peaks fcr establishing retention data and both will compensate for variations in injection volume. An advantage of using a sugar or polyol is that since the internal standard will also be derivatised any variables affecting this procedure should affect test and standard substances equally. Also, test and standard sugars should be equally affected by molecular sieving or water displacement effects that may occur if an ion-exchange

pl~~.vc

-_-.---___~-.

3%QFu1 5% QF-1

TrrIflrroropropyl methyl slllcurte

1%0%17 2% ov-17 3%OV=l7 S%OV-17 10%ov=17

3.8%SE-30 5%SE-30 2% ov.1 3 %0%1 12%OV-1 PlrenylnlerhJll skl%X.wc 3%SE.52

1%SE-30 3% SE-30

_-_Methylslllconc

Slarlonary

. ..-.._

. _ _._ ___

_ me---

TMS Aidononitriic, ncetyl

Methyl TMS Acctyi Butyiboronntc TMS

TMS,ncctyi TMS

Oximc-TMS TMS TMS TMS TMS TMS

Methyl TMS

Urinarydisaccharides Monosnccheridcs and poiyoisin culturemcdin

Swummonosucchnridcs nnd polyois Poiyoisin urine Inositoiin tissues Diwchoridcsin facccs Carbohydrates in biologicalfluids Mnnnitolin urine Mannltolin bioiogicnifluids Sugarawd poiyoisin biologicnifluids

Disucchnridcs in hcces Monosacchnridcs in ccrcbrospinal fluid Myoinositol in bioiogicuifluids Gnlactitolin biologicsifluids Monosaccharides in crythrocytcs Glucoseand poiyoisin urinennd tissues Mow nnddisncchnridcs in urine Sugnrsin,urine Sugarsin bioiogicuifluids Disacchnridcs in bloodnnd urino Monosnccharidcs nndsugnrnicoholsin blood

62 31

;; 55,62,63

53 45 12 67

54

:; 58,64 ti7 62 65

7;73 60

:: 50

12

RCA ___-__--I

.___ .--..-w...-....w......_-__

Derlvnrivc Applicallorr ._-__- -_.._-_---- ._--.- -_.___-_...__- . _- _-..._1.__.__- . . _-._.-- -- -___..-_-__- _.-_.----._--

TABLE1 STATIONARY PHASESUSED FOR SUGARAND SUGARALCOHOLANALYSIS ___-._.___________ .-_____. -.. --. _..___.. - -_.__ __... .._ .- -.___. -_.__-

5

XF-1105

1OlA

SE-52

Caplllwy SE-30

co11ms

Polypropylcrteglycol 5 % WON LB.550

Polymide 3 % poly

2&DEGS

EI/lyk/tcglycol sfK!cClarc 15% EGS

1% EGSS.X

Ethylenewcimte slllcone copoly~ws 3 % ECNSS.M

3.8% CCW-98

v~tlyll,urllyl sillcorte

1 nnd 5 % 0%225

Plrl?tlykcyflrtopPopyl nrclllyl slllcot1c

Cyanoprapyl fllc~llyl Jlli~OllC 3 % SP 2340

2%

Cya/toct/lyl nler/rylslllcotlc 3 % XE=GO

?

Aldononitrilcacctyl Metboximc-TMS TMS

TMS

Acctyl

TMS

TMS

Methoximcncctyl Acctyl TMS

TMS

Acctyl

Acctyl

Alditol formation t trifluoroncctylntion

TMS

Monosacclwidcs and polyols in urine and the lens Monosaccharidesand polyols in scminel fluid Polyolsin ccrebrospinnlfluid

lnositol in tissues

Polyolsin serum nnd urine

Polyolsin biologicalfluids Sugarsin serum and urine Inositol in tissues

Monosaccharidesin blood and urine Polyolsin blood and urine Polyolsin tissues

Sugnrsand polyols in biologicalmatcrinls

Polyolsin ccrcbrospinnlhid and plasma

Gnlactitolin amniotic fluid

Dlsaccharidcsin urine Inositolin tisaucs Sugarsin blood

R 8

41 45,47,53 40 ’ 48

27-29 2 61

48

54

$ Fi

23 18

M.F.LAEER

466

or protein precipitation stage is included_ Hydrocarbons are chemically dissimilar but have the advantage that they do not occur naturahy in biological fluids, and from the great number available one may be easily selected wiffi a retention time which falls within a vacant portion of the chromatogram. The following hydrocarbons and related substanceshave been rrsed: glyceryll-decyl ether”-‘*, ndodecano161, diphenyIbenzene74and diethyiphthalate es&p_ Sugars and polyols that have been used include a-methylghrcose5s~6z, a-methylmannose6’, 3-O-me*hylgluco#, mannitolm-23-81, L-rhamnose-, deoxyriboseG, trehaIOX?, turano+‘, ceIlobiose35,sedoheptulosan hydrate”, perseitoP* and a-methylxylose9. Other substances that have been used in&de uridine~3.Several of these standards do not conform with the requirements listed above. Anomers of 3-Omethylglucose, L-rhamnose, cellobiose and deoxyribose wiil occur unIess double derivatisation procedures are employed. Mannitoi occurs naturally in urine and is not suitable as an internal standard when urinary su_earsare being determined63. Some authors have not used internal standardisation but have relied on careful attention to volumetric technique1°*M~~*‘5-~g_ However, when the two methods of quantitation have been compared greater precision has been achieved with internal standardisatio&=. Peak height or area measurement may form the basis of quantitation if isothermal anaiyses are being performed, but it is customary to measure peak area when temperature pro_mming is used. Partially overlapping peaks may be more satisfactorily quantified by measuring peak he&i@. Sugars that anomerise may be quantified in different ways. It is sometimes possible to select a stationary phase on which the anomers coincide_ Some authors have measured the area of each anomer and added them to give a total peak area for ffiat su_ga?, or measured a combined peak heighp’. Others have reported satisfactory results by using a single anomer for quantitaiion6’. 4.3. Detailed analytical procedure The stages included in a typical analytical procedure are shown in Fig. 1.

aliquotsample 1

add

internals&dard

solution

4 ali!! + (deproteinise) (CkL, +

Fig- 1. Sckme

for 2naiytical proced~_

derikkz c inject The stages in parentheses are Eat always reqk!ired.

Deproteinisatioc of serum samples is necessary to prevent excessive consumption of derivatisation reagents which would resuit in dilution of the sugars with

GLC OF~SUGARS

AND

SUGAR

_ALCOHOI_S

4457

consequent loss of sensitiviQ_ A variety of reagents have been used for protein pre-

cipitation, including the Somoygi te&nique@‘*s, ethanol”,62, perchloric acidi and sulphosdicylic acid62. Separation by ultmfihration has been described in several studies2*S2.54. With Somoygi and ethanol precipitation large volumes result and the fluid must be removed before samples can be derivatised. In many methods additional purilication of samples is undertaken. Urea in high concentrations has been reported to interfere with siIyfation reactionsg5_Some authors remove urea by incubating with ~rease~~~~,others by separating sugars from urea by column chromatography50.It has been shown ffiat the mixed-bed resin Zerolit DM-F (BDH, Poole, Great Britain) reduces urea ConcentrationP and this has been used to treat samples for sugar estimations5.62.Procedures for urinary estimations have been reported which do not include urea removaP3*&. Desalting procedures may be necessary to remove polar constituents which might interfere with derivatisation. Excess protein precipitants and ammonia generated from urea breakdown are usually removed by this technique. A mixed-bed resin~-55-6zor anionic and eationic exchange resins18*s2 may be used. Samples are then dried by evaporation1g*s5.62, lyophilisation~7~54 or desiccatiorP3,following which the residues are derivatised. 5. PERFORMANCE

OF ASSAYS

As for any method, GIG may be assessedfor accuracy (recovery), precision, sensitivity, linearity, interferences and comparison with other techniques. Recoveries are usually determined by adding sugars to the fluid being investigated and estimating the amount present which is then quoted as a percentage of the amount added. These values are usually apparent rather than true recoveries since the amount determined is relative to that of the appropriate standard. When internal standardisation is used recoveries of 90-l 10% are obtained for acetylation and trimethylsiiylation procedures. Monson and Wilkinson30 obtained values of 85-l 13% for mannose determination in plasma using aldononitrile acetylation without internal standardisation Recoveries of 94-104% for glucose determination were obtained by Wells et aLsowho formed TMS ethers using a volumetric technique. With the same method of quantitation, Servo et ~1.‘~found somewhat lower values when polyols were measured as acetyl derivatives in cerebrospinal fluid (75-90x). Gas chromatographic methods are relatively precise, particularly when internal standardisation is used. Intra-assay coefficients of variation of 26% have been obtained for several sugars and polyols in plasma and ~rine?*~‘. Using volumetric techniqtie values of 2-12x have been reported-. One of the principle advantages of GIG is its great sensitivity. Several procedures have been described which will quantify concentrations of l-10 mgfl (refs. 43,46,55 and 62). Unrelated substances,in addition to other sugars, may have a similar retention time to the sugar being estimated. Specificity may be checked by analysing on several stationary phases and by using alternative derivatisation procedures, since it is unlikely that an interfering substance will co-chromatograph with several different methods. If necessary, the structural nature of a peak may be characterised by combiied gas chromatography-mass spectrometry.

468

M.F.LAKER

Glucose determinations by GLC have been compared with several ghrcose oxidase-based methods, correlation coefikients of O-98 and 0.99 being obtained~_ A similar correlation coefficient was found when Iactniose estimations by quantitative paper chromatography and GLC were compared‘l . These resuits show that the accuracy of GLC methods for sugar determinations is comparable with these alternative techniques_ 6_ CONCLUSIONS

Many GLC procedures are available for measuring sugars and sugar aIcohoIs in biological t%rids_They should be considered where it is necessary to measure these substances simuhaneousiy, to resolve complex mixtures, or to quantity low concentrations7. ACKNOWLEDGEMENTS

It is a pleasure to acknowledge helpful advice from Drs. W. D. Mitchell and I_ S. Menzies, and from Mr. J. N. Mount. 8. SUMMARY

Gas-Iiqtud chromatographic procedures for measuring sugar and pofyol concentrations in bioIogical fluids are reviewed. Such methods require the preparation of derivatives such as methyl ethers, trimethylsilyl ethers or acetyl esters. Prior to derivatisation samples must be deproteinised and dried. Complex mixtures of sugars and sugar alcohols may be resolved_ Quantttative analyses are precise, sensitive and Iinear. If internal standardisation is used recoveries approaching 100% are obtained_ REFERENCES 1 R. L. Jo&y, K. S. Warren, C. D. Scott, J. L. Jainchilland AU_L. Friedman, Amer. J_ Clin. Parh., 53 (1970) 793. 2 P. Starset, C. Stokke and E. Jellum, J. Chrumarogr.,145 (1978) 351. 3 D. E. Vance and C. C. Sweeley, Methods Med. Res_, 12 (1970) 123. 4 J. R CJamp, Biochem. Sot. Trans., 5 (1977) 1693. 5 J_ R Clamp, Biochem. Sot_ Symp., 40 (1974) 3. 6 T_ E. Acree, R_ S. Shallenbxger and L. R. Mattick. Carbohyd. Res., 6 (1968) 4998. 7 A. G. Mcim~es, D. H. Bail, E P_ Cooper and C. T_ Bishop, J. Chrummgr_, I (1958) 556. 8 H. W. Kixher, Anal. Chenz., 32 (1960) 1103. 9 C. T. Bishop, Merh& Biochem_Ad.., 10 (1%2) 1. 10 Yu. S. Ovodov and E. V. Evtashenko, J. Chroin~togr., 31 (1967) 527. 11 S_ Hhkomo< J_ Biochem_(Tokyo). 55 (1964) 205 12 R. M. Thompson, R. D. Co&t. M_ R. G&k and J. F_ Fitzgerald, C&n. Chti. Actu, 84 (1978) 185. 13 S_ W_ Gunner. J. K. W_ Jones and M. B. Perry, Cizn_J_ Cfrem.. 39 (l%l) 1892. 14 W. J. A. VandenHeuvel and E. C. Homing, Bh%ent_ Biophys. Ret. Comnum.. 4 (1961) 399. 15 H_ G. Jones and M. B. Peny, Can. I_ Chem., 40 (1%2) 1339. 16 J_ S_ !$avardeker. J. H. Sloneker and A. J-, Ad. Chem, 37 (1965) 1602. 17 J_ M. Oades, J_ Ciucxna~ogr_,28 (1%7) 246. 18 E_ Pi&&en, Ch. Chti_ Acra, 38 (1972) 221.

GLC OF SUGARS

AND

SUGAR

ALCGHGLS

469

19 E. Pi&&en and C. Servo, Cfin. Cfuin. AC&Z,44 (1973) 437. 20 C. Servo, J. Pa!0 and E. Pitk&~en, Acca iVewoC Scmrd., 56 (1977) lU4. 21 R G&&man, H.-C. Curtius and I. SchnelIer, Erp. Eye Res., 6 (1967) 1. 22 H. S&wind, F_ Scharbert. R. Schmidtand R. Katterman, i. C&z. Chem. Clin. Biochem., 16 (1978) 145. 23 D. Murphy and C. A. Pennock, C&t. Cfiim. Acta, 42 (1972) 67. 24 J. T. Allen, M_ Gillett, J. B. Holton, G. S. King and B. R. Pettit, Lancet, i (1980) 603. 25 R. Varma, R. S. Varma and A. H. Wardi, i. C?zrunz~~ogr., 77 (1973) 222. 26 R. Varma and R. S. Varma, J. Chromnrogr., 128 (1976) 45. 27 J. Szafi-anek, C. D. PfaKenberger and E. C. Horning, Anal. Serf., 6 (1973) 479. 28 C. D. Pfaffenkger, 5. Szafbiek, M. G. Horning and E. C. Homing, Anal. Biuchem., 63 (1975) 501. 29 C. D. PfhfF&hrger, J. Szafi-anekand E. C. Homing, J. Chromomgr., 126 (1976) 535. 30 T. P_ Mcmson and K. P. Wil!&son, Clin. Chem., 25 (1979) 1384. 31 G. J. Ma&s, T. M. Y_ Lui and L.-F. L. Wen, Anal, Biochem., 99 (1979) 365. 32 E. J. Bourne, C. E. M. Tatlow and J. C. Tatlow, J. Chem. Sot., (1950) 1367. 33 M. Vi&as, H.-L Jan, G. Eoussac and M.-C. Bonnard, TetrahedronLetr., (1966) 1441. 34 J. Shapiia, T. Putkey, A. Furst and G_ A. McCasland, Cprbohyd. Res.. 25 (1972) 535. 35 H. Nakamma and Z. Tamura, Clin. Ctiim. AC~Q, 39 (1972) 367_ Chem Phorm_ BuB_. 15 (1967) 246. 36 Z. Taxnun and T. I&, 37 J. P. Zmetta, W. C. Breckenridge and G. Vincendon, J. Chromotogr., 69 (1972) 291. 38 J. E. Sullivan and L. R. S&ewe, J. Chromotogr. Sci., 15 (1977) 196. 39 C. C. Sweeley, R. Bentley, M. h&k& and W. W. Wells, J. Amer. Gem. Sot., 85 (1963) 2497. 40 W. W. Wells, T. Chin and B. Weber, CL+. Chim. Acta, 10 (1964) 352. 41 W. W. Wells, T. A. Pittman and H. J. We&, Anal. Biuchem.. 10 (1965) 450. 42 R. Quan-Ma and W. W. Wells, Biochem. Biophys. Res. Commzm., 20 (1965) 486. 43 T. Bhatti and J. R. Clamp, Cf&_ Chim. Acto, 22 (1968) 563. 44 J_ de NeeF, Clin. Chim. Acta, 26 (1969) 485. 45 W. R. Sherman and M. A. Stewart, Biochem. Biophys Res. Commun., 22 (1966) 492. 46 M. A. Stewart, W. R. Sherman, M. M. Kurien, G. I. Moonsammy and M. Wisgerhof, J. Nemothem., 14 (1967) 1057. 47 W. R. Sherman, M. A. Stewart, M. M. Kurien and S. L. Goodwin, Biochim. Biophys Acta, 158 (1968) 197. 48 K. Narmni, M. Arita. M. Kitagawa. A_ K umazawa and T. Tsumita, Jap. J. Exp. Med., 39 (1969) 399. 49 L. M. Leain, A. Szeinberg and E. Lepkifker. Clin. Chim. Acta, 45 (1973) 361. 50 R. S. Ctements and W_ R. Starnes, Biachem. Med., 12 (1975) 280. 51 S. Amundsson, A. I. Braude and C. F. Davis, App. Microbioi., 28 (1974) 298_ 52 D. J. Hesf and D. J. Galton, Clin. Chim. Acfa, 63 (1975) 41. 53 E. Pit&ten and K. Sahlstriim, Ann. Exp. Med. Fenn., 46 (1968) 295. 54 J. F. Aloia, 1. Lob. Clin. &fed., 82 (1973) 809. 55 M. F. Laker and J. N. Mount, Clin. Chem., 26 (1980) Ml. 56 G. Peterson, Carbohyd. Res_, 33 (1974) 47. 57 C. F. Poole, in K. Blau and G. King (Editors), Ha&book of Derivativesfor Chrumafography, Heyden, London, 1978,Ch. 4, p_ 157. 58 W_ C. Butts and R. L. Jolley. Cfin. Chem., 16 (1970) 722. 59 K. Osaka, M. Kuboyama, I. ?Kiyosawa, T. Suzaki 2nd M. Itoh, J. Nun+_ScL Vitamhzol~, 21 (1975) 129. 60 T. VanderMeuIen and R. A. Hudson, Anal. Biuchem., 53 (1973) 639. 61 S. L. Smith, M. Novotay and E. L. Weber, Clin. Chem., 24 (1978) 545. 62 M. F. Laker, 1. Chromomgr_, 163 (1979) 9. 63 M. F. laker and W. G. Gunn, C&I. Chim. Acta, 96 (1979) 265. 64 E. C. Horning and M. G. Horning, Methods Med. Res., 12 (1970) 396. 65 E. C. Horning and M. G. Horning, Metha% Med. Res., 12 (1970) 400. 66 H. van Nicolai. & Egge and F. Zilliken, 2. Nafnr$wsch. C, 30 (1975) 451. 67 W. W. WelIs, Me&&s Med. Res., 12 (1970) 115. 68 P. E. Reid, B. Donaklson, D. W. Secret and B. Bradford, f. Chronuztogr., 47 (1970) 199.

470 69 70 71 72 73 74

75 76 77 78 79 80 81

52

M. F. LAJSR D. V. PhiNips and A. E. Smith, Anal. Brbchem_, 54 (1973) 95 A. Erzcson, i. Hansen ati t. Dalgaarcl, ,4.&. Biochcm..,86 (1978) 552. M. F. Laker, .U.D. lXesr& University of Landon, 1979.p_ 237. F. Eisenberg. Carbuhytf. Res_. 19 (1971) 135. D. L. Sondak. f_ Phurm- Sci., 64 (1975) 129. C. Sthweickhardt, J_ Clin. cirem- Clin. Bioc&m_, 16 (1978) 675_

D. G. Pritchard and W. Nb&xmeier~ f. C~runmmgr_. 152 (1978) 487_ S. Pate!, J. Rivlin. T_ Samueelson,O_ A_ S&mm and H. ZoiIinger, Heh-_ C&M_ Acia, Sl(1968) 169. W. W_ Wells. T. A. Pittman and T_ I. Egan, I. Bill Cbem, 239 (1964) 3192. W. W. Wells, J. BiaI. Chem.. 240 (1965) 1002. 1. J. Egan and W. W. Wells, /fnzer.J. Dis: ChiH, 111 (1966) ooo, C. T_ Bishop, A&an. CarbohytL C&m., 19 (!964) 9.5. J. X. Clasp, T. Bhatti and R. E. Chambers, Meniodr Bkxiiem. Am!., 19 (1971) 229. C. E. D&$&h, E. C. Homing, Mz. G. Horning, K. L. Knoxand K. Yargezr,Bmhenr..l., 1Oi (1966)

792. 83 F. D. Gauche:, G. Wagner and K. H. B&&x,

J. C&n. Chem. Chin. Brbchem., 9 (1971) 25. 84 C. J. W. Bmoks and J. A_ Zabkiewicz.in C. H. G--y and A. L. Bachamch (Editors), hormones in BZo&, Vol. 2, Academic Press, London, 1967,Ch. 3, p. 64. . 85 W. W. Wells, C. C. SweeIeyand R. Ben&y, in H. A. Szymanski(Editor), i3&medicaIAppCcafi~ns of Gas Chromurography, Plenum E%ess, New York, 1964, p. 208. 86 I. S. M&es, personal comtitio~_ .%’ C. A. Pcxmock,D. Murphy, J. Se&n and K. J. Longdon, Clin. C&n. .-&to. 48 (1973) 1933.